ABSTRACT A key limitation for using direct semiconductor detection for positron emission tomography (PET) is the relatively slow drift times of the charge carriers which leads to poor coincidence timing. We propose that Cerenkov light can be used as a fast timing signal in semiconductors such as TlBr. Cerenkov light is emitted quasi-instantaneously when a gamma-ray generates a photoelectron and has the potential to provide a signal with a time-of-flight precision even for semiconductor detectors. Cerenkov imaging has been proposed just a few years ago as a new modality and the first experimental results of Cerenkov light detection with BGO have been published recently, showing a coincidence time resolution of 270 ps. Cerenkov light in TlBr will be more intense than in BGO due to its higher index of refraction. The utilization of Cerenkov light as a fast timing signal in semiconductor detectors has not been reported yet. TlBr is an optimum detector for PET in terms of efficiency and energy resolution, however, its timing response compromises its overall performance. Using the Cerenkov light as a fast signal could provide TlBr with a time-of-flight timing performance and convert it into one of the most complete detectors for PET. The goal of this proposal is to demonstrate and characterize the detection of Cerenkov light in TlBr. The specific aims address understanding of the generation and transmission of Cerenkov light in TlBr, and the detection of this light using TlBr slabs coupled to photodetectors. The three specific aims of this project are 1) the measurement of visible light transmittance in TlBr, 2) the simulation of Cerenkov light generation in TlBr, and 3) the detection of Cerenkov light generated in TlBr and measurement of timing resolution. The experimental measurements will be used to validate the simulation and to obtain a toolkit able to predict reliably the Cerenkov light emitted by a TlBr slab depending on its dimensions and to extrapolate it to other semiconductor materials. These specific aims will be sufficient to prove the feasibility of using Cerenkov light as a fast signal in TlBr and to provide the semiconductor detector community with a simulation toolkit that facilitates the production of TlBr detectors optimized to detect Cerenkov light.